computational fluid dynamics simulation of the combustion...

COMPUTATIONAL FLUID DYNAMICS SIMULATION OF THECOMBUSTION PROCESS, EMISSION FORMATION AND THE FLOW

FIELD IN AN IN-DIRECT INJECTION DIESEL ENGINE

by

Ramin BARZEGAR a*, Sina SHAFEE b, and Shahram KHALILARYA c

a Young Researchers Club, Parsabad Mogan Branch, Islamic Azad University, Parsabad, Iranb Department of Mechanical Engineering, Faculty of Natural and Applied Sciences,

Middle East Technical University, Ankara, Turkeyc Mechanical Engineering Department, Urmia University, Urmia, Iran

Original scientific paperDOI: 10.2298/TSCI111218108B

In the present paper, the combustion process and emission formation in the Lister8.1 in-direct injection diesel engine have been investigated using a computationalfluid dynamics code. The utilized model includes detailed spray atomization, mix-ture formation and distribution model which enable modeling the combustion pro-cess in spray/wall and spray/swirl interactions along with flow configurations. Theanalysis considers both part load and full load states. The global properties arepresented separately resolved for the swirl chamber (pre-chamber) and the mainchamber. The results of model verify the fact that the equal amount of the fuel isburned in the main and pre-chamber at full load state while at part load the major-ity of the fuel is burned in the main chamber. Also, it is shown that the adherence offuel spray on the pre-chamber walls is due to formation of a stagnation zone whichprevents quick spray evaporation and plays an important role in the increase ofsoot mass fractions at this zone at full load conditions. The simulation results, suchas the mean in-cylinder pressure, heat release rate and exhaust emissions are com-pared with the experimental data and show good agreement. This work also demon-strates the usefulness of multi-dimensional modeling for complex chamber geome-tries, such as in in-direct injection diesel engines, to gain more insight into the flowfield, combustion process, and emission formation.

Key words: in-direct injection diesel engine, computational fluid dynamics, flowfield, combustion, NOx, soot

Introduction

In a very competitive world, improvement of engine performance has become an im-

portant issue for automotive manufacturers. In order to improve the engine performance, de-

tailed studies are carried out on combustion process and emission formation. There are two com-

mon methods in the field of engine research; experimental study and mathematical simulations

(thermo dynamical modeling, computational fluid dynamics (CFD) modeling, etc.).

The experimental methods obtain principal and valuable information, but they require

costly equipment and are mostly time consuming. The other approach which has become a dom-

inant factor in the engine research in the recent years is the application multi-dimensional CFD

codes which can gain insight into the in-cylinder flow field and combustion process by solving

the governing flow field and species transport equations [1, 2].

Barzegar, R., et al.: Computational Fluid Dynamics Simulation of the Combustion Process, ...THERMAL SCIENCE: Year 2013, Vol. 17, No. 1, pp. 11-23 11

* Corresponding author; e-mail: [email protected]

Diesel engines are preferable to gasoline engines because of their high thermodynamic

performance and low HC, CO, and NOx emissions. The main problem in diesel combustion

though is the soot formation, or smoke, which is produced when there is insufficient air for com-

plete combustion [3-6]. In an in-direct injection (IDI) diesel engine, the combustion chamber is

divided into two major parts: the pre-chamber and the main chamber which are linked by a

throat. In an IDI diesel engine, fuel injects into the pre-combustion chamber and air is pushed

through the narrow passage on the compression stroke and becomes turbulent within the

pre-chamber. This narrow passage speeds up the expanding gases more. The pre-chamber ap-

proximately includes 50% of the combustion volume when the piston is at top dead centre

(TDC). This geometry adds an additional difficulty to those to deal with in the direct injection

(DI) combustion chambers. Generally speaking, IDI engines have lower performance than DI

engines, which is because of their intense heat loss in swirl chamber throat [7]. 3-D modeling

takes into account the interactions between different phenomena including turbulent flow,

spray, combustion and naturally the geometry of combustion chamber. It allows a precise inves-

tigation of the problem as it provides all required properties at any one point within the combus-

tion chamber at any one time. When applied to IDI engines, these models have to address spe-

cific problems linked to the flow unsteadiness, high Reynolds numbers involved, and the

complex variable geometry of the solid boundaries.

As a consequence, the CFD simulations usually take long calculation time requiring

high memory capacity, but these problems have been partially solved as a result of the signifi-

cant improvement in power and speed of modern computers in recent years. In other words, sim-

ulation of the combustion system by means of computer modeling makes it possible to explore

combustion regimes that may be difficult and/or expensive to achieve with experiments [8].

There are many reports on experimental studies regarding the IDI diesel engines in the

literature such as study of alternative fuels, improving of combustion process and emission re-

duction [9-13]. Hotta et al. [14] worked on an experimental research to reduce the particulate

matter (PM) from IDI diesel engine and examined the reducing mechanisms using an optically

accessible engine at Toyota central research and development laboratories, Inc. It is now com-

monly admitted that the design of IDI combustion chambers has to rely more on fundamental

knowledge of local aspects which require multi-dimensional simulation. Many fundamental as-

pects concerning CFD simulation of IDI engines have been discussed earlier by Pinchon [15].

3-D modeling of combustion process and soot formation in an IDI diesel engine have also been

studied by Zellat et al. [16] using KIVA code. Strauss and Schweimer [17], at Volkswagen AG,

had studied the combustion and pollutant formation processes in a 1.9 l IDI diesel engine using

SPEED CFD code for the part load and full load conditions.

Study of the relevant literature shows that only few attempts have been done in order to

use the multi-dimensional modeling in IDI diesel engine research up to now. At present work,

the AVL Fire CFD code has been used to predict and investigate the flow field, combustion pro-

cess, performance parameters and emissions in the Lister 8.1 IDI diesel engine at full and part

loads. Taking into account the rather complicated nature of the IDI diesel engines due to the

continuous mass and energy exchange between the two chambers, we have decided to make an

investigation on in-direct injection diesel engine.

Initial and boundary conditions

Calculations are carried out on a closed system from intake valve closure (IVC) at

165° CA bTDC to exhaust valve opening (EVO) at 180° CA aTDC. Figure 1 shows the numeri-

cal grid, which is designed to model the geometry of the engine and contains a maximum of

Barzegar, R., et al.: Computational Fluid Dynamics Simulation of the Combustion Process, ...12 THERMAL SCIENCE: Year 2013, Vol. 17, No. 1, pp. 11-23

42200 cells at 165° CA bTDC. The present

resolution was found to give adequately grid

independent results. The diesel fuel is in-

jected via a single hole injector in pre-cham-

ber as shown in fig. 2. Initial conditions for

pressure and temperature in the combustion

chamber are 86 kPa and 384 K, respectively,

and swirl ratio is assumed to be at quiescent

condition. The engine speed under study is

730 rpm. for both full load and part load

states. All boundary temperatures were as-

sumed to be constant throughout the simula-

tion, but allowed to vary at combustion

chamber surfaces.

Model description

The specifications of the Lister 8.1 IDI

diesel engine are listed on tab. 1. The flow

was solved from IVC to EVO using the AVL

FIRE CFD code [18]. The AVL CFD code is

powerful software which is used by re-

searches in internal combustion engine field

and car manufacturers to investigate different

aspects of combustion and emission forma-

tion in engines. This code solves the com-

pressible, turbulent, 3-D transient conserva-

tion equations for reacting multi-component

gas mixtures with the flow dynamics of an

evaporating liquid spray. The turbulent flow

within the combustion chamber is simulated

using the RNG k-e turbulence model, modi-

fied for variable-density engine flows [19].


Figure 2. Spray and injector co-ordination atpre-chamber

Table 1. Specifications of Lister 8.1 IDI dieselengine

Cycle type Four stroke

Number of cylinders 1

Injection type IDI

Cylinder bore 114.1 mm

Stroke 139.7 mm

L/R (rod/stroke ratio) 2

Displacement volume 1.43 litre

Compression ratio 17.5:1

Vpre-chamber/VTDC 0.7

Full load injected mass 6.4336e-5 kg per cycle

Part load injected mass 3.2009e-5 kg per cycle

Max power on 850 rpm 8 hp

Max power on 650 rpm 6 hp

Injection pressure 91.7 kg/cm3

Start injection timing 20° bTDC

Nozzle diameter at holecenter

0.003 m

Number of nozzle holes 1

Nozzle outer diameter 0.0003 m

Spray cone angle 10°

Valve timing

IVO = 5° bTDC

IVC = 15° ABDC

EVO = 55° BBDC

EVC = 15° aTDC

Figure 1. (a) Mesh for the Lister 8.1 IDI dieselengine; (b) top view of the mesh

The standard WAVE model [20] is used for

the primary and secondary atomization

modeling of the resulting droplets. At this

model the growth of an initial perturbation

on a liquid surface is linked to its wave

length and other physical and dynamical

parameters of the injected fuel at the flow

domain. Drop parcels are injected with

characteristic size equal to the Nozzle exit

diameter (blob injection). The injection

rate profiles at full load state and 50% load

are shown in fig. 3.

The Dukowicz model is applied for treating the heat up and evaporation of the drop-

lets, which is described in [21]. This model assumes a uniform droplet temperature. In addition,

the droplet temperature change rate is determined by the heat balance, which states that the heat

convection from the gas to the droplet either heats up the droplet or supplies heat for vaporiza-

tion. A Stochastic dispersion model was employed to take the effect of interaction between the

particles and the turbulent eddies into account by adding a fluctuating velocity to the mean gas

velocity [18].

The spray/wall interaction model used in this simulation was based on the spray/wall im-

pingement model [22] which has previously been utilized by authors to investigate effects of

fuel injection mode and spray impingement on diesel combustion [1, 2]. The Shell auto- -igni-

tion model was used for modeling of the auto ignition [23]. In this generic mechanism, six ge-

neric species for hydrocarbon fuel, oxidizer, total radical pool, branching agent, intermediate

species, and products were involved. In addition the important stages of auto ignition such as

initiation, propagation, branching and termination were presented by generalized reactions, de-

scribed in [18, 23]. The Eddy break-up model (EBU) based on the turbulent mixing was used for

modeling of the combustion in the combustion chamber [18]. NOx formation is modeled by the

Zeldovich mechanism and soot formation is modeled by Kennedy, Hiroyasu, and Magnussen

mechanism [24].

Results and discussion

The calculations are carried out for the single cylinder Lister 8.1 IDI diesel engine and the

operating conditions are full and 50% load at constant speed of 730 rpm. Figures 4 and 5 show

the comparison of computed and measured [25] mean in-cylinder pressure and heat release rate

(HRR,) respectively, for both load cases. The good agreement between measured and predicted

data especially during the compression and expansion strokes verifies the model. Both premixed

and diffusion combustion phases are computed as well. Comparing these figures also shows the

effect of load on the heat release rate and in-cylinder pressure. The measured HRR curve is de-

rived from the first law analysis of procured in-cylinder pressure data as:

d

d

d

d

d

d

Qp

VV

P

q

g

g q g q�

��

�1

1

1(1)

In this equation, P and V are in-cylinder pressure and volume vs. crank angle q, and g =

= 1.35. The results presented in the figures are global (cylinder averaged) quantities as a func-

tion of time (crank angle). In order to study the part load condition, only the fuel injection timing

and amount of injected mass were changed and the rest of the initial conditions were unaffected.


Figure 3. Part load and full load injection histories

The peak pressures discrepancy between experiment and computation is less than 0.2%. In-

creasing load to full load mode causes in-cylinder peak pressure to increase to 50.2 bar from

42.3 bar and ignition delay decreased to 7.9 crank angle degree (CAD) from 10.7 CAD with re-

spect to 50% load. At full load operation, due to injection of fuel in later parts of cycle and also

longer injection duration, much of the fuel is burned in diffusion phase. Figure 4 also shows the

quantity of computational and experimental results for start of combustion (SOC) and ignition

delay (ID) CAD in both cases. The discrepancies of SOC or ID between computation and exper-

iment at part load and full load operation are as small as 0.8° and 0.1° crank angle, respectively.

Figure 6 shows the development of the burnt mass fraction. There is a classic result found in re-


Figure 4. Comparison of measured [23] and calculated pressure at 730 rpm for (a) full load and (b) partload

Figure 5. Comparison of experimental and calculated heat release rate at 730 rpm for (a) full load and(b) part load

Figure 6. Calculated burn fuel mass fraction at 730 rpm for (a) full load and (b) part load

lation between the values for the main and

the pre-chamber burnt fuel mass. It can be

seen that the SOC is earlier in the full load

state compared with the part load case in the

pre-chamber. Also at full load state, the

computed burn fuel mass fraction and heat

release rate are equality shared between the

main chamber and pre-chamber. This shows

that simulations have correctly represented a

typical feature at part load states in IDI com-

bustion with fuel burning mostly in the main

chamber. Figure 7 also indicates the accu-

mulated heat release rate at full and part load

operations.

The mean in-cylinder temperature histories are presented region-resolved for the swirl,

main and total in-cylinder chamber in fig. 8. The peak temperature values in the swirl chamber,

main chamber, and in total occur, respectively, at 1394, 1612, and 1389 K for part load and 1788,

1966, and 1827 K for full load operation, respectively. With the increase of load at constant speed

(730 rpm) from, the peak temperature in swirl, main and total in-cylinder chambers increase by

394, 354, and 438 K, respectively. The increase in load means an increase in the equivalence ratio

since the amount of injected fuel is increased for constant air mass in the pre-chamber. This tends

to increase the air temperatures and thus increases the chances of forming a flammable mixture for

pre-flame reactions to take place.

This reduces the ID in the pre-chamber. Also at full load operation, because of the lon-

ger injection duration, combustion period is longer than the part load operation, hence, tempera-

tures in the swirl, main and total in-cylinder chambers at full load are higher than the 50% part

load operation in the expansion stroke. Also in the IDI diesel engine, since the fuel spray is in-

jected into an auxiliary chamber such as the swirl chamber, most of the fuel spray will remain in

the pre-chamber, especially at full load operation. The impinged fuel gradually evaporates and

burns. Therefore combustion period in pre-chamber is long at full load operation, fig. 8(a).

Figure 9 represents vectors of the velocity field at various crank angles in horizontal

cuts from main combustion chamber and planes across the connecting throat. It was found out

that the maximum velocities obtained here were lower than the corresponding data in the litera-


Figure 7. Accumulated heat release rate at730 rpm for full and part load operations

Figure 8. Comparison of experimental and calculated heat release rate at 730 rpm for (a) full load and(b) part load

ture [7, 14-16] due to large areas at throat section. The maximum velocity for full load operation

is then 56 m/s at 10° CA aTDC which is a result of start of combustion occurring in pre-chamber

and maximum mass outflow. Figure 8 also indicates that the swirl generated during the com-

pression stroke (20° CA bTDC at fig. 9) becomes gradually weaker due to opposite flow in glow

plug. At 60° CA aTDC, velocities get dominated by the evacuation of pre-chamber. The flow in

throat changes its direction at the beginning of combustion and after TDC the flow in the

pre-chamber is strongly influenced by the fuel spray. The top view corresponding to 20° CA

aTDC depicts the formation of two vortices which are created by the deceleration of the axial

penetration of the mixture. In the front view at 40° CA aTDC, the gas from the pre-chamber

reaches the opposite side of cylinder. This leads to the formation of two large eddies each occu-

pying a half of the main chamber and staying centered with respect to the two half of the bowl.

At 60° CA aTDC, these eddies are bigger. Flow field at part load operation is similar to full load,

whereas that maximum velocity is 61 m/s.


Figure 9. Velocity vector plots for full load operation

In fig. 10, contours of the fuel spray and vapor saturation are presented from start of in-

jection at 10° CA bTDC to 35° CA aTDC. The velocity flow field shown in fig. 9 explains why

the spray conserves initial direction imposed by the injector. The fuel vapor-air mixing in the

pre-chamber is better in the full load case than the part load condition due to the existence of vor-

texes at the full load operation. Also, impingement of fuel spray against the chamber wall is a re-

sult of less interaction between flow field and spray. It is observed that increased engine load

yields in more fuel vaporization and spray droplets in the pre-chamber at the beginning of com-

bustion. This is probably the origin of the much higher soot formation in the pre-chamber at full

load operation. The hard spray impingement causes fuel adhesion on the wall near spray-wall

impinging point. This adherent fuel is not quickly evaporated and the formed fuel vapor is

hardly carried out of this area which is a stagnation zone due to the chamber shape. Thus it is

possible that the rich fuel-air mixture stagnates in this zone under the condition of high tempera-

ture and insufficient oxygen to form the dense soot cloud.

Performance parameters

Table 2 shows the variation of performance parameters vs. load operation, compared

with the experimental data [25]. By comparison with the experimental results, it can be seen that

the model predictions have a good accuracy. Indicated work per cycle was calculated from the

cylinder pressure and piston displacement as:

W P V� � dq

q

1

2

(2)

where q1, and q2 are the start and end of the valve-closed period, respectively, (i. e. IVC = 15°

aBDC and EVO = 55° bBDC). The indicated power per cylinder and indicated mean effective

pressure were related to the indicated work per cycle by:


Figure 10. Spray and vapor saturation contours

PWN

n�

60000(3)

IMEPW

V�

d

(4)

where n = 2 is the number of crank revolutions for each power stroke per cylinder, N – the engine

speed, and Vd – the volume displacement. The brake specific fuel consumption (BSFC) was de-

fined as:

BSFCm

P� f

b

(5)

In eq. 2, the work was only integrated as part of the compression and expansion strokes

and the pumping work was not taken into account. Therefore, the power and ISFC analyses can

only be viewed as being qualitative rather than quantitative in this study.

Table 2. Performance parameters predictions and compared with experiments

Part load mode Full load mode

CFD model Experimental CFD model Experimental

Indicated power[kW]

3.006 – 5.773 –

Brake power [kW] 2.555 2.535 4.907 5.064

�mfuel [gs–1] 0.196 0.196 0.391 0.391

BMEP [bar] 2.98 2.96 5.72 5.90

BSFC [gkW–1h–1] 276 277 287 278

Emission formation results analysis

In the following section, the production of NOx and soot as main emissions in IDI diesel

engines are discussed. Exhaust NOx values are compared and verified with experimental data

for both cases. Comparison shows that the predicted values are in good agreement with mea-

sured data for the full load operation case while at part load operation there is diversity between

results. As the fuel injection is retarded, the time available for air/fuel mixing and combustion is

reduced, and the consumption rate of the fuel decreases. The fuel consumption rate becomes

much slower with very late injection, resulting in significant increases in the unburned fuel

amount.

As indicated by fig. 11, under medium and high load conditions, the main cause of the

exhaust smoke can be both the fuel adhesion to the chamber wall and stagnation of rich fuel-air

mixture. Also it can be seen that the NOx formation is intensified in areas with equivalence ratio

close to 1 and the temperature higher than 2000 K. In addition to this, the area which the equiva-

lence ratio is higher than 3 and the temperature is approximately between 1600 K and 2000 K

yields in greater soot concentration. A local soot-NOx trade-off is evident in these contour plots

where the NOx and soot formation occur on opposite sides of the high temperature region [26].

The in-cylinder temperature histories reveal that the flame started in pre-chamber and then in-

vades a large portion of the main-chamber very quickly. It is observed that the beginning of the



Figure 11. NOx, soot, temperature, oxygen, and vapor mass fractions contour plots at full load operation(from left to right) for TDC, 10°, 20°, 30°, and 40° CA aTDC (from top to bottom)(for color image see journal web site)

mass transfer between pre-chamber and the main chamber is around TDC. The development of

the temperature field between 20° and 30° CA aTDC shows that the axial and the radial penetra-

tions of the flame front are almost equal. Thus, the flame front reaches the lateral cylinder wall

and the cylinder wall opposite to the pre-chamber at approx. the same time. The fast decrease of

the oxygen concentration in main and pre-chambers correspond to a large high temperature zone

(over 2000 K at 10° CA aTDC).

The region resolved global NOx formation of for the two cases are given in figs. 12(a)

and 12(b). NOx formation starts off at about 17° CA for full load and 18° CA for part load opera-

tion state after the start of injection. The initial increase in the global NOx formation follows the

global temperature trend (also on fig. 8). Also it is observed that NOx formation in main chamber

is higher compared to that in pre-chamber especially at part load operation. Finally, figs. 12(c)

and 12(d) present the evolution of the soot in the chamber and in the pre-chamber for both load

conditions. One could notice that at part load, the maximum soot mass is observed in the main.

However, soot oxidation is fast the exhaust gas is already attained at about 60 CA aTDC. In fact,

because of the availability of sufficient oxygen, the turbulent soot oxidation diminishes the soot

concentrations to 0.1 g/kWh at about 110° CA aTDC. At high loads, soot quantities in the

pre-chamber is much bigger than main chamber due to stagnation region forming in glow plug

and also soot is produced in regions of high fuel concentrations, where cold fuel is injected into

areas of hot gases. The soot is then oxidized again in the leaner regions of the flame, so that most

of the soot is already consumed in zones close to the stoichiometric. Soot quantities in the

pre-chamber vanish due to its oxidation and convection to the main chamber.


Figure 12. Comparison of measured [23] and calculated exhaust emission at 730 rpm for full load andpart load operation

Conclusions

In this paper, the mixture formation, combustion processes, pollutant formation, flow

field, and the performance parameters within Lister 8.1 IDI diesel engine were simulated at full

load and part load operation conditions. The global properties were presented resolved for the

swirl and main chamber separately. As for the interpretation of the results, it was shown that the

weak deflection of the fuel jet was because of the low swirl level in the pre-chamber due to the

presence of the glow plug. Results for calculated pressure, exhaust emissions and performance

parameters were compared with the corresponding experimental data and show good agree-

ment. Such verification between the experimental and computed results gives confidence in the

model prediction, and suggests that the model may be used at future works. The CFD simula-

tions have also shown that they greatly improve the understanding and to facilitate the analysis

of the combustion and pollutant formation processes in the multi-chamber diesel engines.

References

[1] Jafarmadar, S., et al., Numerical Investigation of the Effect of Fuel Injection Mode on Spray-Wall Im-pingement and Combustion Process in a Direct Injection Diesel Engine at Full Load State, Thermal Sci-ence, 14 (2010), 4, pp. 1039-1049, DOI: 10.2298/TSCI10041039J

[2] Jafarmadar, S., et al., Modeling the Effect of Spray/Wall Impingement on Combustion Process and Emis-sion in a DI Diesel Engine, Thermal Science, 13 (2009), 3, pp. 23-33, DOI: 10.2298/TSCI0903023J

[3] Heywood, J. B., Internal Combustion Engine Fundamental, McGraw Hill, New York, USA, 1988[4] Benson, R. S., Whitehouse, N. D., Internal Combustion Engines, Pergamon Press, Oxford, USA, 1979[5] Ferguson, C. R., Internal Combustion Engines, John Wiley and Sons, New York, USA, 1986[6] Obert, E. F., Internal Combustion Engines and Air Pollution, Intext Education Publ., New York, USA,

1993[7] Patterson, D. J., Henein, N. A., Emissions from Combustion Engines and Their Control, Science Publ.,

Ann Aebor, Mich., USA, 1972[8] Uludogan, A., et al., Modeling the Effect of Engine Speed on the Combustion Process and Emissions in a

DI Diesel Engine, SAE paper 962056, 1996[9] Sanli, A., et al., The Influence of Engine Speed and Load on the Heat Transfer between Gases and In-Cyl-

inder Walls at Fired and Motored Conditions of an IDI Diesel Engine, Applied Thermal Eng., 28 (2008),11-12, pp. 1395-1404

[10] Canakci, M., et al., Combustion Analysis of Preheated Crude Sunflower Oil in an IDI Diesel Engine, Bio-mass and Bioenergy, 33 (2009), 5, pp. 760-767

[11] Selim, M. Y. E., et al., Combustion of Jojoba Methyl Ester in An Indirect Injection Diesel Engine, Renew-able Energy, 28 (2003), 9, pp. 1401-1420


Nomenclature

�mf – fuel consumption rate, [kgs–1]N – engine speed, [rpm]P – mean cylinder pressure, [kPa]Pb – brake power, [kW]Pi – indicated pressure, [kPa]Q – heat released in combustion chamber,

– [kJ]V – cylinder volume, [m3]Vd – displacement volume[m3]W – work per a cycle, [kJ]

Greek symbol

q – crank angle, [deg.]

Acronyms

aTDC – after top dead centreBMEP – brake mean effective pressureBMEP – brake mean effective pressureBSFC – brake specific fuel consumptionbTDC – before top dead centreCFD – computational fluid dynamicsDI – direct injectionEBU – eddy brake upEVO – exhaust valve openingIDI – in-direct injectionIVC – inlet valve closureSOC – start of combustion

[12] Celikten, I., An Experimental Investigation of the Effect of the Injection Pressure on Engine Performanceand Exhaust Emission in Indirect Injection Diesel Engines, Applied Thermal Eng., 23 (2003), 16, pp.2051-2060

[13] Parlak, A., et al., The Effects of Injection Timing on NOx Emissions of a Low Heat Rejection Indirect Die-sel Injection Engine, Applied Thermal Eng., 25 (2005), 17-18, pp. 3042-3052

[14] Hotta, Y., et al., Combustion Improvement for Reducing Exhaust Emissions in IDI Diesel Engine, SAEpaper 980503, 1998

[15] Pinchon, P., Three Dimensional Modeling of Combustion in a Pre-Chamber Diesel Engine, SAE paper890666, 1989

[16] Zellat, M., et al., Three Dimensional Modeling of Combustion and Soot Formation in an Indirect InjectionDiesel Engine, SAE paper 900254, 1990

[17] Strauss, T. S., Schweimer, G. W., Combustion in a Swirl Chamber Diesel Engine Simulation by Computa-tion of Fluid Dynamics, SAE paper 950280, 1995

[18] ***, AVL FIRE user manual, Ver. 8.5, 2006[19] Han, Z., Reitz, R. D., Turbulence Modeling of Internal Combustion Engines Using RNG k-e Models,

Combust Sci. and Tech., 106 (1995), 4-6, pp. 267-295[20] Liu, A. B., Reitz, R. D., Modeling the Effects of Drop Drag and Break-up on Fuel Sprays, SAE paper

930072, 1993[21] Dukowicz, J. K., Quasi-Steady Droplet Change in the Presence of Convection, Informal Report Los

Alamos Scientific Laboratory, Los Alamos, N. Mex., USA, LA7997-MS[22] Naber, J. D., Reitz, R. D., Modeling Engine Spray/Wall Impingement, SAE paper 880107, 1988[23] Halstead, M., et al., The Auto Ignition of Hydrocarbon Fueled at High Temperatures and Pressures – Fit-

ting of a Mathematical Model, J. Combust Flame, 30 (1977), pp. 45-60[24] Patterson, M. A., et al., Modeling the Effects of Fuel Injection Characteristics on Diesel Engine Soot and

NOx Emissions, SAE paper 940523, 1994[25] Mohammahi, A., Ignition of Dual Fuel Engines by Using Free Radicals Existing in EGR Gases, Ph. D. the-

sis, Faculty of Mechanical Engineering, Tabriz University, Tabriz, Iran 2008[26] Carsten, B., Mixture Formation in Internal Combustion Engines, Springer Publications, Berlin, 2006

Paper submitted: December 18, 2011Paper revised: April 8, 2012Paper accepted: April 20, 2012


computational fluid dynamics simulation of the combustion...

Documents